Gyro's Picoammeter; output ESD protection

[ecclectrics]

Hi All,

motivation of the topic comes from planning to build the little gadget described in these two threads:

https://www.eevblog.com/forum/projects/picoammeter-design/
https://www.eevblog.com/forum/beginners/static-control-requirements-for-picoamp-measurements-using-ucurrent-gold/

but is of more general interest. Schematic from the original thread is attached for reference. While I do understand the main workings of the circuit, I wonder about the output protection which as mentioned in the threads "was added as an afterthought". Issues I am trying to sort out are the following:

1) Really the only realistic threat I can imagine to appear there is ESD, or am I overlooking something?.

2) I have hardly ever seen a circuit with ESD protection on the output (at the input yes, at the output really only overcurrent and kickback in case of inductive loads). So is ESD protection at an output really needed? Or are IC output structures by themselves typically strong enough to survive?

3) Now I wouldn't mind putting in a few etra components just to be on the safe side, but then I would want them to really work in case of an ESD event. By IEC-61000-4-2 one would have to expect peak currents up to several 10A depending on the test level, albeit only for durations less than ~100ns. I wonder wether the Zeners would survive this without anything between them and the ESD source?

4) Zener datasheets show thermal resistance only for much longer pulses. Roughly extrapolating these values and doing the thermal calculations gives permissible peak currents of <2A for the BZX55 series and <5A for the ZPD series as examples. A spice simulation shows (despite of all the involved approximations) that adding a current limiting resistor of ~1...3.3k would be needed to bring the current down to that range. With a small shunt capacitor of ~100 ... 300p the resistor value could be reduced (of course stability of the opamp would need to be reconsidered). Is a careful selection of diode type or the inclusion of further components necessary to make the protection work or am I overly concerned?

If anyone with more experience on thse specific issues can help, I would greatly appreciate!

Best regards,
Roland

[Kleinstein]

The outputs are usually relatively robust by themslf. So for DIY circuit one rarely cares about extra ESD protection there. Commercial products may or may not have extr aprotection. The specs may be only for the less demanding human body case, though an output connection may indeed be an example where the machine case may apply.

The peak current is usually quite short and even the normal zeners may survive quite some ESD pulses.
If the maximum current is an issue, there are special protection diodes (transsorb, transsil) that can withstand higher power (e.g. 1.5KE  gives a peak current of some 200A for 10 ms).
A relatively common way for protection are diodes towards the supply. The ESD pulse would than charge the fitler capacitors.

[Gyro]

^ What Kleinstein said.

It really was an afterthought - I didn't want to mess around replacing the LMC662 if I did something silly. With 4mm output jacks, there is always the danger of them accidentally getting hooked to the wrong thing on an untidy bench. Worst case, zeners and diodes normally fail short circuit on overcurrent (unless physically blown apart) so can be treated as sacrificial.

[ecclectrics]

Thank you Kleinstein and Gyro, that helps!

If I may ask one further question for potential future applications:

What do you think about moving the 1k resistors between the output and the diodes as sketched in the image? If the Zener capacitance would impede the stability of the opamp, it could be reduced by the arrangement shown with two additional small signal diodes. I think a frequent problem with the TVS diodes mentioned by Kleinstein is their much higher capacitance compared with ordinary Zeners.

Looking forward to hearing your opinions

[Gyro]

Having some resistance between the surge protection device (be it Zener or TVS) and the outside world helps greatly in reducing peak discharge current and is always worth doing where possible - It's actually what I should have done on the Picoameter!

I don't think the zener capacitance would affect the opamp stability, yes TVSs do have higher capacitances depending on type - you can get low capacitance ones for fast signals. Putting two in parallel like that could help in some low signal level circumstances but as soon as the signal voltage exceeds the forward drop of one of the series diodes, the signal is exposed to the TVS capacitance (I think).

If you're looking at the Picoammeter specifically, I don't think there is any need to go beyond simple back-to-back Zeners once you have 1k in series with the accessible outputs.


EDIT: Happy now? 

It is still stable as the additional capacitance on the opamp output is tiny compared to the existing strays....

[ecclectrics]

Wow Gyro, you even tried the modification on your circuit, that's amazing! I am always happy if I can learn something

Now I need to make a plan for the construction and find some time to spend at the workbench ...

Thanks again,
Roland